SOURCE-LINKED INTELLIGENCE

CREDO: Variance-Guided Rubric Evolution for Replay-Corrected Credit Assignment

arXiv · AI, language, vision and robotics · article · Sep 21, 2026 · UTC

Long-horizon language agents receive sparse terminal feedback, while intermediate rubrics provide structured but potentially misspecified assessments of progress. In resettable training environments, counterfactual continuation rollouts can measure local credit, but exhaustive replay is costly. We propose Credo, a framework that couples evolving semantic rubrics with selective, execution-based credit correction. A frozen judge maps visible transitions to rubric features, and a credit head predicts the change in expected terminal reward associated with the realized transition. Independently sam

Read original source ↗ Open in workspace

recordType

paper

region

Global